Identification of Hookworm DAF-16/FOXO Response Elements and Direct Gene Targets

Background The infective stage of the parasitic nematode hookworm is developmentally arrested in the environment and needs to infect a specific host to complete its life cycle. The canine hookworm (Ancylostoma caninum) is an excellent model for investigating human hookworm infections. The transcription factor of A. caninum, Ac-DAF-16, which has a characteristic fork head or “winged helix” DNA binding domain (DBD), has been implicated in the resumption of hookworm development in the host. However, the precise roles of Ac-DAF-16 in hookworm parasitism and its downstream targets are unknown. In the present study, we combined molecular techniques and bioinformatics to identify a group of Ac-DAF-16 binding sites and target genes. Methodology/Principal Findings The DNA binding domain of Ac-DAF-16 was used to select genomic fragments by in vitro genomic selection. Twenty four bound genomic fragments were analyzed for the presence of the DAF-16 family binding element (DBE) and possible alternative Ac-DAF-16 bind motifs. The 22 genes linked to these genomic fragments were identified using bioinformatics tools and defined as candidate direct gene targets of Ac-DAF-16. Their developmental stage-specific expression patterns were examined. Also, a new putative DAF-16 binding element was identified. Conclusions/Significance Our results show that Ac-DAF-16 is involved in diverse biological processes throughout hookworm development. Further investigation of these target genes will provide insights into the molecular basis by which Ac-DAF-16 regulates its downstream gene network in hookworm infection.


Introduction
Many parasitic nematodes, including hookworms, infect the definitive host as developmentally arrested third-stage larvae (L3). The L3 is analogous to the dauer stage of the free-living nematode Caenorhabditis elegans in many biological aspects [1,2,3]. The FOXO-family forkhead transcription factor DAF-16 mediates dauer formation of C. elegans in response to cues indicating poor environmental conditions. When conditions improve, DAF-16 is negatively regulated by an insulin-like signaling (ILS) pathway that culminates in transport of phosphorylated DAF-16 out of the nucleus [4,5,6,7,8,9,10]. The primary protein structure of DAF-16 contains a conserved forkhead or ''winged helix'' DNA binding domain (DBD) with three major a-helices and two large wing-like loops [11,12,13]. Orthologs of DAF-16 have been recently characterized in the parasitic nematodes Ancylostoma caninum (Ac-DAF- 16), Strongyloides stercoralis (Ss-DAF- 16), and Haemonchus contortus (Hc-DAF-16) [14,15,16,17]. Heterologous rescue of C. elegans daf-16 mutants [15,16] and reporter assays in mammalian cells [17,18] indicate that parasitic nematode DAF-16 orthologs play similar regulatory roles during development, providing further support for the use of dauer exit as a model to investigate the molecular events of infection and successful establishment of a parasitic relationship with the host [1].
Murine DAF-16/FOXO was shown to bind an 8-bp consensus DAF-16 family member binding element (DBE) in vitro [19]. Several approaches have since been used to identify DAF-16 target genes in C. elegans, with the results suggesting that DAF-16 is recruited to a large number of promoters to modulate the expression of genes involved in development, metabolism, stress responses, and longevity [20,21,22,23].
Our lab has been focusing on the infectious process of hookworms, one of the most common infectious diseases in tropical and subtropical countries, causing anemia and malnutrition in almost a billion people [24]. The canine species A. caninum is a commonly used model for investigation of human hookworm infections. The DAF-16 ortholog from A. caninum (Ac-DAF-16) was shown to be transcriptionally active and capable of interacting with a hookworm 14-3-3 protein, suggesting a critical role in gene expression associated with hookworm L3 development and the transition to parasitism [17,18]. Given the functional conservation between dauer recovery and hookworm infection, there is considerable interest in the transcriptional outputs of DAF-16 in hookworm and their function in parasitic development. Dissecting the hookworm DAF-16 downstream effector network will have important implications in the development of new intervention strategies for hookworm and other nematode infections.
The present study utilizes in vitro genomic selection, a technique built on the concept of systematic evolution of ligands by exponential enrichment (SELEX), where natural genomic sequences are used as a source for selection and amplification [25]. A combination of in vitro genomic selection and subsequent cloning has been developed as a powerful method to identify naturally occurring DNA-binding sites in a genomic context and provide a foundation for investigation of the in vivo targets of DNA-binding proteins [26,27,28]. We employed the in vitro genomic selection strategy using Ac-DAF-16 DBD to screen digested hookworm genomic DNA, and identified high-affinity binding sites in the hookworm genome and potential Ac-DAF-16 gene targets. Finally, the expression profile of the Ac-DAF-16 related transcripts was determined by examining cDNAs from four developmental stages of A. caninum.

Ac-DAF-16 DNA binding domain (DBD) expression and purification
Amino acid sequence alignment of different FOXO proteins revealed that the DBD is approximately 100 amino acids-long, with a conserved N-terminal region, and a divergent, but arginine/lysine-rich C-terminal region (Fig. 1). Several crystal structures for FOXO transcription factor DBDs have been solved, and reveal that the structural basis for FOXO protein recognition of DNA comes from direct base-specific contacts as well as phosphate contacts between DNA molecules and critical C-terminal arginine/lysine amino acid residues in the FOXO DBD [29,30]. Based on the information from those crystal structures and sequence alignment between different DAF16/FOXO molecules, the DBD of Ac-DAF-16 was defined to start at amino acid Asn214 and end at amino acid Asp314.
Properly functioning recombinant hookworm DAF-16 DBD peptide was required for the in vitro genomic selection technique. A fragment of 303 bp corresponding to Ac-DAF-16 DBD (aa 214-314), and a 249 bp fragment corresponding to a truncated Ac-DAF-16 DBD (aa 220-302) lacking the arginine/lysine-rich Cterminal region (DAc-DAF-16 DBD) were cloned and expressed ( Fig. 2A). The calculated molecular masses for the coding regions of these two constructs are 14047.4 Da (pET28a-Ac-DAF-16 DBD) and 12510.6 Da (pET28a-DAc-DAF-16 DBD). As expected, bands corresponding to the predicted molecular weights (14 kDa for rAc-DAF-16 DBD, and 12 kDa for rD Ac-DAF-16 DBD) were detected by Coomassie Blue staining (Fig. 2B). Immunoblots using an anti-His tag (C-term) antibody indicated that the bands were present only in E. coli cultures that had been transformed with the expression constructs and induced with IPTG (Fig. 2C, Lanes 3 and 4).
Ac-DAF-16 was previously shown to bind to the consensus DBE sequence and initiate reporter gene transcription [17]. Our pulldown assay results (Fig. 3) indicated that rAc-DAF-16 DBD, but not rD Ac-DAF-16 DBD, recognizes and binds strongly to the conserved DBE, indicating that the arginine/lysine -rich section at the C-terminus of DBD is critical for its binding activity.

In vitro genomic selection
The in vitro genomic selection enabled unbiased identification of transcription factor binding sites in the absence of influence from chromatin and other cofactors [25]. Immobilized rAc-DAF-16 DBD was first prepared by binding to anti-FLAG M2 affinity Figure 1. Amino acid sequences of DNA binding domains from selected FOXO transcription factors were aligned using CLUSTAL W software on the San Diego Supercomputer Center Biology Workbench server (http://workbench.SDSC.edu) and displayed using BOXSHADE3.21 software located on Swiss EMBnet server (http://www.ch.embnet.org). Identical amino acids are in red type, and conserved amino acids in blue. C-terminal arginine and lysine residues are shaded. The DBDs are from the following species: AcDAF16 (Ancylostoma caninum accession ACD85816); CeDAF16 (Caenorhabditis elegans, AAB84390); Foxo3, (Mus musculus, AAH19532); Foxo1 (Homo sapiens, AAH70065); Foxo4 (Homo sapiens, AAI06762). doi:10.1371/journal.pone.0012289.g001 matrix, which was confirmed by silver staining (Fig. 4 A) and Western blot (Fig. 4 B), and used to screen BfuCI digested A. caninum total genomic DNA (Fig. 4C). Discrete bands appeared progressively over the four rounds of binding and PCR amplification (Fig. 4D, lanes 2 to 5), suggesting preferential amplification of particular genomic fragments. Cloning the amplified DNA fragments from the fourth round resulted in a total of 311 clones (117 pBluescript KS+ constructs and 194 pGEM-T Easy constructs), and high-quality sequences were obtained for 274 of them. The length ranged from 100 bp to 300 bp. Sequence analysis showed that 198 sequences contained low-complexity microsatellite regions, which were characterized by the presence of the repetitive trinucleotide ''GTT'' or its reverse complement ''AAC'' with a repeat number of 5 to 15. The remaining 76 sequences represented 25 distinct genomic fragments, and 24 of them were successfully mapped to A. caninum genomic sequences. A single fragment remained unidentified after extensive sequence similarity searching against nucleotide databases at Genebank, suggesting that it might by located in a section of the A. caninum genome that has not been sequenced. Five of the 25 fragments were overrepresented, and the remaining 20 fragments were recovered once or twice ( Table 1). The sequences of the distinct genomic fragments were submitted to the Genome Survey Sequence database (dbGSS) at the NCBI, and the accession numbers reported in Table S1.
A control selection was performed to control for non-specific binding to the antibody-bead matrix. A total of 211 clones were picked, and 190 high quality sequences were obtained. Sequence analysis showed that anti-FLAG M2 affinity matrix bound to a different set of hookworm genome fragments (Table S2). Among these sequences, only one was shared with the DAF-16 DBDselected fragments, indicating that selection using the DBD  Biotinylated dsDBE (Bio-DBE) was incubated with purified Ac-DAF-16 DBD or DAc-DAF-16 DBD, and peptide/oligonucleotides complexes were pulled down with streptavidin conjugated Sepharose beads. Precipitated DBD/oligo complexes were separated by SDS-PAGE and blotted to PVDF membranes for Western blotting using an anti-His (Cterm) antibody. Bio-random represents a biotinylated oligomer of random sequence. doi:10.1371/journal.pone.0012289.g003 generated specific genomic fragments associated with DAF-16 binding elements.

Motif analysis of rAc-DAF-16 DBD bound hookworm genome fragments
The consensus DBE is an 8-bp oligonucleotide, with the sequence 59-TTGTTTAC-39 [19]. The core sequence of DBE is TGTT, and a single base-pair replacement in this sequence might significantly weaken its interaction with forkhead proteins. Therefore, only a single mismatch outside of the DBE core sequence was allowed when the 24 selectively bound fragments were inspected for DBEs. Using this criterion, 13 fragments contained the DBE or its reverse counterpart (Table 1), and two contained two copies of the DBE arranged as direct repeats separated by variable length of nucleotides. The remaining 11 genomic fragments were analyzed further, and a new overrepresented sequence, 59-GAC/GAA/TG-39, was found, occurring 17 times in 9 of those fragments (Fig. 5a, Table 1). Genomic fragments selected by the anti-FLAG M2 control were also analyzed for over-represented motifs. One element (59-AGGAA-GAG-39) was found in 36% of the control fragments, but only 5% (5/103) of the fragments contained a DBE-like element (Table S2), indicating that the DBE and over-represented GAC/GAA/TG motif were bound specifically by rAc-DAF-16 DBD.
The newly identified motif was tested for its ability to bind to Ac-DAF-16 DBD. Oligos containing the four most abundant motifs (59-GACAAG-39, 59-GACATG-39, 59-GAGAAG-39, 59-GA-GATG-39) were tested in a streptavidin bead pull-down assay. The oligos failed to pull down rAc-DAF-16 DBD (data not shown). One possible explanation for this result was that additional flanking sequence might be required for efficient binding. To test this possibility, biotinylated T7 primer was used to amplify representative cloned genomic fragments, which were tested for their ability to bind the DAF-16 DBD. As shown in Figure 5b, rAc-DAF-16 DBD strongly bound to positively selected genomic fragments that contained a DBE (fragments 4.6 and 2.23) or the newly identified motif (fragments 3.23 and 3.28), but not to the anti-FLAG M2 affinity matrix selected (control) genomic fragments. This suggests that sequence flanking the element is required for in vitro binding, and that the newly identified motif specifically binds to the Ac-DAF-16 DBD when located in the proper context.

Genes or gene clusters linked to the recovered rAc-DAF-16 DBD bound fragments
Transcription factor binding sites are typically linked to their direct target genes. To identify the genes or gene clusters linked to the rAc-DAF-16 DBD bound fragments, approximately 6 kb of the A. caninum genome scaffold sequences [39], (Mitreva, unpublished) containing the fragments at the center were searched against stagespecific hookworm cDNA databases. Nineteen of the 24 genome scaffolds retrieved about 5000 cDNAs from different developmental stages, one scaffold matched an rRNA region, and the other four scaffolds either had poor sequence alignments or failed to retrieve any cDNAs ( Table 2, Table S3).
Clustering the retrieved cDNAs yielded 22 transcript contigs. One rAc-DAF-16 bound fragment was allowed to be linked to one or two transcript contigs. Sequence alignment between those transcripts and genome scaffolds revealed that rAc-DAF-16 bound fragments resided in coding regions, introns, or 39 untranslated regions (39-UTR) (Table S3). However, not all the relative locations, especially the upstream ones, could be detected by this method due to incomplete annotation of the hookworm genome. Transcript sequences were searched against protein databases in Genbank and Gene Ontology Consortium for prediction of their function based on homology. The results classified rAc-DAF-16 linked transcripts into two groups. The first contained transcripts homologous to known or hypothetical proteins, and displayed a variety of putative biological processes such as intraand extracellular signaling (PKA, NLP-2, SRT-42), metabolism (CPT-II, ABC transporter class F, phosphate transporter), development (cuticlin), and transcription/translation (SNR-3); the second contained the transcripts without any identified homologs across available protein databases and therefore represent putative A. caninum specific genes.
The recent studies of Wang et al [31] generated 1.5 million A. caninum cDNAs from four developmental stages of A. caninum: infective L3 larva (L3), serum stimulated (activated) L3 larva (aL3), adult male (M), and adult female (F), covering approximately 93% of the A. caninum transcriptome. This dataset allowed construction of a stage-specific digital expression profile for many transcripts [31]. Examining the available expression pattern profiles indicated that five rAc-DAF-16 linked transcripts (PKA, contig00807, contig25170, contig12656, and contig08715) were significantly up-regulated and three (contig40879, contig04080, and phosphate transporter) were significantly down-regulated by serum stimulation (table 2, Table S3), which is hypothesized to mimic the very early events of hookworm infections [1]. Three rAc-DAF-16 linked transcripts (contig00807, contig25170, and NLP-2) are larval specific, and therefore turned off during development from L3 to adult. A major difference between parasitic hookworms and freeliving C. elegans is that hookworms are dioecious, and the differential digital expression profile indicated that nine rAc-DAF-16 linked transcripts are gender-specific (PKA, contig12656, contig40879, cuticlin1-17, CPT, contig04080, phosphate transporter, contig08715, and contig44862). Therefore, the DAF-16 linked transcripts identified here are involved in multiple biological processes during at least 4 hookworm developmental stages.

Discussion
During infection, hookworm L3 resume developmental programs that had been arrested during the environmental stage, resulting in the successful establishment of a parasitic relationship with their host. Most of the molecular events associated with this ''transition to parasitism'' are still unknown. Dauer recovery of the free-living nematode C. elegans has long been used as a model for investigating the mechanisms of this transition to parasitism due to the biological similarities between the dauer stage and infective L3 of hookworms [1,2,3]. Specifically, the conserved insulin signaling pathway, which is required for C. elegans to exit the dauer stage in response to the improving environmental conditions, is also involved in the hookworm infection process [32,33,34]. Studies using inhibitors have shown that the ILS pathway is involved in hookworm larval activation, and the hookworm ortholog of the central transcription factor in this signaling pathway, DAF-16, has been identified and characterized [17,35]. Furthermore, Ac-DAF-16 was transcriptionally active and interacted with hookworm 14-3-3 protein in a phosphorylation-status dependent way in cultured mammalian cell [17,18]. DAF-16 is a FOXO transcription factor that is negatively regulated by ILS [4]. It functions in numerous biological processes, including metabolism, life span, stress responses, and dauer formation and exit, by mediating downstream gene expression in response to environmental and nutritional conditions [36]. Many of the downstream targets have been identified in C. elegans, and provide insights into the mechanism by which DAF-16 mediates multiple phenotypes [20,21,22,23]. While many of these mechanisms are conserved in hookworms, the life cycles of parasitic nematodes differ significantly from that of C. elegans, and at least some of the DAF-16 outputs are likely to mediate processes specific to nematode parasitism. Therefore, fundamental mechanistic questions about hookworm infection can be addressed by the identification of DAF-16 binding sites and direct gene targets, and how those downstream effectors are coordinated in hookworm development.
To begin identifying Ac-DAF-16 target genes, we used an in vitro genomic selection technique to enrich for genomic fragments that were bound by the recombinant Ac-DAF-16 DBD. In vitro genomic selection with transcription factors differs from other affinity-based strategies because it uses the native genomic background, and therefore direct transcription factor targets are identified without influences from complex cellular environments [37]. The cyclical strategy selects binding sites with relatively high affinity and reduces indirect or nonspecific binding. In the present study, a group of genomic fragments with high affinity for rAc-DAF-16 DBD were enriched, as indicated by higher-than-expected occurrence of some genomic fragments. Sequence analysis indicated that more than half of those fragments contained a reported consensus DBE or DBE-like element, and these elements were present in four of the five most represented DBD bound genomic fragments. The DBE or DBE-like elements occurred as single or multiple copies, and the element orientations and nucleotide spacers between the elements varied. As C. elegans DAF-16 and its mammalian homologs bind to this consensus DBE, the presence of the DBE in Ac-DAF-16 selected hookworm genome fragments suggests a conserved function for hookworm DAF-16. However, nearly half of the fragments we isolated did not contain a canonical DBE or DBE-like sequence, suggesting that the DAF-16 DBD bound to a previously unknown binding element. Further analysis of these fragments identified an over-represented 6-bp element, 59-GAC/GAA/TG -39. The Ac-DAF-16 DBD bound to amplicons containing this element, but not individual oligos, suggesting that the new element requires surrounding sequence for DBD binding. Our results do not rule out the presence of additional DAF-16 binding elements in the hookworm genome. It is not unusual that a single transcription factor has variable bind sites [38]. In any case, we have identified a new, previously unreported DAF-16 DBD binding element (GAC/GAA/TG) in hookworms, supporting a role for Ac-DAF-16 in multiple, perhaps novel, hookworm-specific biological processes.
The availability of the draft genome of A. caninum and comprehensive expression data enabled a detailed analysis of the DBD-bound genes. Twenty four of 25 rAc-DAF-16 DBD selectively bound fragments were confirmed as A. caninum, and their proximal genomic regions were analyzed for coding regions. A total of 22 transcripts within a 6 kb range surrounding rAc-DAF-16 DBD bound fragments were identified as Ac-DAF-16 primary target genes. The use of a 6 kb search range was based on C. elegans transcription factor binding site analyses [20,34]. However, not all the genomic sequence hits in the present study returned transcript contigs within this range. Therefore, an extended search might be necessary to identify them, as A. caninum has a larger genome than C. elegans [39] and the A. caninum genome project is still underway.
Functions of the proteins encoded by the transcripts were predicted based on their homology to known or hypothetical proteins. This analysis suggested that DAF-16 is involved in a variety of biological processes. For example, PKA is a conserved serine/ threonine kinase, activated by second messenger cAMP, and converts various extracellular signals into intracellular processes [40]. Serpentine receptors are G protein-coupled transmembrane receptors that play important roles in C. elegans chemoreception [41]. The NPL protein is critical for synaptic transmission between neurons [42], and membrane transporters are involved in transport of a wide variety of substrates across extra-and intracellular membranes [43,44]. Cuticlins comprise the insoluble, high-order material in nematode cuticle and determine the developmental morphology and mobility of the worms [45]. Small nuclear riboproteins (SNR) are a part of RNA post-transcriptional modification machinery [46]. Among these genes, the serpentine receptor and ABC transporter were also reported in previous studies of DAF-16 targets in C. elegans, indicating that some pathways downstream of DAF-16 are conserved in free living and parasitic nematodes [20,22,23]. Additionally, some rAc-DAF-16 DBD linked gene transcripts failed to match any homologs by exhaustive search of the available databases, and were defined as A. caninum specific. These are of particular interest, as they are absent from C. elegans and consequently might be involved in parasitism. Furthermore, several of these molecules could be envisioned functioning during the transition to parasitism based on homology and expression apttern.
Combining the transcriptomic and genome sequences revealed the relative location of Ac-DAF-16 binding sites to their linked genes in the A. caninum genome. However, this method is biased towards identification of Ac-DAF-16 binding sites located between exons due to incompleteness of hookworm genome annotation and the inability to identify 59 ends and promoter sequences. Nonetheless, all 10 fragments that could be definitively linked to a gene had binding elements in introns or downstream sequences, suggesting that they may be regulated differently from genes with DBEs in the promoter. In C. elegans, DAF-16 target genes containing DBEs located downstream of the start codon were more likely to be negatively regulated by DAF-16 [47]. Characterizing intergenic Ac-DAF-16 binding sites will depend on further information about the genomic structure for the corresponding genes. The present study surprisingly indicates that Ac-DAF-16 binding sites, unlike C. elegans DAF-16 binding sites, reside at variable locations relative to the gene transcripts [23], suggesting that Ac-DAF-16 might regulate expression of genes with diverse functions and exert its action through different mechanisms.
In the absence of functional information for most of the identified rAc-DAF-16 linked genes, evidence of differential expression is the most important source for prioritizing future investigations. Using the extensive cDNA dataset available for A. caninum, the expression profile for the gene transcripts have been quantitatively analyzed by comparing the frequency of EST occurrence in the different cDNA libraries. Examination of the expression patterns for the identified transcript contigs in the present study suggests that Ac-DAF-16 regulates gene expression in all hookworm developmental stages studied, including exit from developmentally arrested infective L3 stage, maturation to adults, and sexual differentiation in adults.
Using the affinity-based in vitro genomic selection procedure, we have shown for the first time that Ac-DAF-16 directly binds to response elements in the hookworm genome. The relative location of Ac-DAF-16 bound elements to the linked genes is variable, with an apparent bias towards downstream locations. The Ac-DAF-16 direct target candidates that were identified include both conserved and A. caninum specific genes, and will be subject to future functional investigations. With more comprehensive screening such as chromatin immunoprecipitation and improved A. caninum genome and transcriptome data, more Ac-DAF-16 downstream targets will be detected. Subsequent manipulation of these genes may lead to novel avenues for intervention in the hookworm life cycle.

Ac-DAF-16 DBD cloning, expression and purification
To clone the Ac-DAF-16 DBD, a fragment containing the Cterminally FLAG-tagged DBD (corresponding to amino acids 214-314) was amplified from a cDNA clone of Ac-DAF-16 isoform b, pCMV4-daf16 [17]. The specific forward primer (DAF16-DD-E-FLAG-FN: 59-GATACCATGGGCAA TGCGTGGGGTAA-TCTC-39, containing restriction site NcoI, underlined) and reverse primer (DAF16-DD-E-FLAG-RH: 59-GATCAAGCTTATCGT-CG TCATCCTTGTAGTC CAAGGTGGAGGCTCG AAC-39, containing restriction site HindIII, and the FLAG Tag, italics), were incubated with the template in a PCR. The cycling conditions were 2 min at 95uC; followed by 35 cycles of 1 min at 95uC, 1 min at 55uC, 1 min at 72uC and a final extension for 6 min at 72uC. The purified amplicon was digested with NcoI and HindIII, ligated into expression vector pET28a (Novagen) cut with the same restriction enzymes, and transformed into Escherichia coli DH5a competent cells. The expression construct contained both an inframe FLAG tag and an in-frame hexahistidine (His) tag at Cterminus of the Ac-DAF-16 DBD, as confirmed by DNA sequencing (Nevada Genomics Center, Reno, NV).
The resulting plasmid DNA (pET28a-Ac-DAF-16 DBD/ FLAG/His) was transformed into E. coli Rosetta (DE3) competent cells (Strategene), and expression was induced by addition of 1 mM isopropyl-b-D-thio-galacto-pyranoside (IPTG) to log-phase bacterial culture at 37uC for 4 hrs. An aliquot of E. coli Rosetta (DE3) cells were removed from the culture prior to induction to serve as a pre-induction control, and second aliquot from a culture grown under the same conditions except in the absence of IPTG served as an un-induced control. Induced bacterial cells and control bacterial cells were collected by centrifugation at 5000 rpm for 15 mins. After cell lysis by sonication in the presence of protease inhibitors (Pierce, Thermo Scientific, Rockford, IL), the expressed recombinant Ac-DAF-16 DBD (rAc-DAF-16 DBD) was affinity-purified by Ni-NTA resin (Qiagen, Valencia, CA). Purified rAc-DAF-16 DBD peptide was fractionated by SDS-PAGE eletrophoresis through 4-20% Tris-glycine pre-cast Novex gradient gels (Invitrogen, Carlsbad, CA), and examined by staining with Coomassie Brilliant Blue R250 (Sigma, St. Louis, MO) and Western Blot with anti-His (C-term) antibody (Invitrogen). The concentration of the expressed peptide was determined by the Micro BCA Protein Assay (Thermo Scientific).

Ac-DAF-16 DBD Binding Assay
The DNA binding ability of rAc-DAF-16 DBD was tested by streptavidin bead pull-down assay using double stranded DNA as described previously [17]. Briefly, 59 end biotin-labeled forward oligonucleotides and their unlabeled complements for the DBE or predicted motif elements were synthesized (IDT, Coralville, IA). Forward and reverse oligonucleotides were annealed to form double-stranded (ds) DNA and 100 pmoles were used in the pulldown assays. To label selected genomic fragments, 59 end biotinlabeled T7 primer was synthesized and used to amplify the fragment from a plasmid clone by PCR. The amplicons were purified and 2 pmoles were used in the pull-down assays. Biotinlabeled dsDNA was incubated with 200 ng of Ac-DAF-16 DBD in binding buffer (10 mM Tris pH 7.5, 50 mM KCl, 1 mM DTT, 2.5% glycerol, 5 mM MgCl 2 , 50 ng/ mL poly (dI NdC), 0.05% Nonidet P-40) at 4uC for 2 h. After incubation, 25 ml of 30% Streptavidin-Sepharose (GE Healthcare) slurry equilibrated with TNE 50 buffer (10 mM Tris, 50 mM NaCl, 1 mM EDTA, pH 7.5)+0.1% Nonidet P-40 were added to the mixture and incubated for 2 h at 4uC. Beads were collected by centrifugation at 2500 g and washed twice with TNE 100 buffer (10 mM Tris, 100 mM NaCl, 1 mM EDTA pH 7.5) +0.1% Nonidet P-40. The bound proteins were separated on a 4-20% SDS-PAGE and Western blotted with anti-FLAG antibody (Sigma).

Hookworm genomic DNA preparation
The Baltimore strain of A. caninum (US National Parasite Collection #100655.00) was maintained in beagles as previously described [48]. The George Washington University Medical Center Institutional Animal Care and Use Committee approved this study (protocol #A147). Infective L3 larvae were recovered from coproculture by a modified Baermann technique and stored in buffer BU (50 mM Na 2 HPO 4 , 22 mM KH 2 PO 4 ,70 mM NaCl, pH 6.8) [49] at room temperature until used.
Eighty thousand hookworm L3 were frozen in liquid nitrogen and ground to a fine powder using a pre-chilled mortar and pestle. Following physical disruption of the worms, hookworm genomic DNA was isolated using the Wizard SV Genomic DNA Purification System (Promega) according to the manufacture's instructions. Briefly, digestion solution master mix containing RNase A (75 mg/mL) and proteinase K (1.5 mg/mL) was added to the homogenate and the sample was incubated at 55uC overnight. The sample was then mixed with lysis buffer, applied to the minicolumn assembly, and eluted with buffer TE (10 mM Tris.HCl, 1 mM EDTA, pH 8.0). Fifteen micrograms of A. caninum genomic DNA were digested with 48 units of BfuCI restriction endonuclease (NEB) overnight and stored at 220uC. The quality of DNA was examined by 0.8% agarose gel electrophoresis and the concentration was determined using a NanoDrop ND-1000 spectrophotometer.

In vitro Genomic Selection
Thirty microliters of anti-FLAG M2 affinity matrix (Sigma) were rinsed with TBS buffer (10 mM Tris HCl, pH 7.4, 150 mM NaCl, 0.1 mM EDTA) three times. Two micrograms of purified rAc-DAF-16 DBD/FLAG/His were immobilized on the prepared matrix by incubation for 2 hours at 4uC in 150 mL of TBS buffer. The matrix was then washed with TBS buffer containing 1 M NaCl, followed by two washes with the binding buffer (10 mM Tris at pH 7.5, 2.5% glycerol, 10 mM MgCl 2 , 50 mM KCl, 1 mM dithiothreitol (DTT), 0.05% NP-40). All washes were performed at 4uC for 5 min on a nutating mixer. Immobilized Ac-DAF-16 DBD was then incubated with 5 mg of BfuCI digested A. caninum genomic DNA, in 150 mL of binding buffer containing 50 ng/ mL poly (dINdC) for 30 min at room temperature on a nutating mixer. To control for non-specific binding and subsequent PCR enrichment of spurious genomic fragments, a control selection using only the anti-FLAG M2 affinity matrix was performed under identical conditions. Protein-DNA matrix complexes were washed with 150 mL binding buffer for 5 min on a nutating mixer at 4uC, followed by a wash with binding buffer containing 250 mM KCl and a third wash containing 500 mM KCl. Washing conditions were optimized with the consensus DBE, and 500 mM KCl was shown to be sufficient to remove most of the unspecific binding. Ac-DAF-16 DBD-bound DNA was eluted from the matrix with 250 mL of the binding buffer containing 1 M KCl (10 min with 150 mL followed by 15 min with 100 mL) on a nutating mixer. Two hundred and fifty microliters of TE (10 mM Tris HCl, pH 8.0, 1 mM EDTA) were then added to the eluate. The eluted DNA was phenol/chloroform-extracted and ethanol-precipitated in the presence of 5 mg of glycogen. The DNA pellet was re-dissolved in 12 mL H 2 O. The following oligonucleotides, which contain the NotI restriction site, were synthesized: forward NotIBfuCI (59-AAAAGGGGCGGCCGC-39) and reverse NotIB-fuCI (59-GATCGCGGCCGCC CCTTTT-39), and annealed to form double-stranded (ds) NotIBfuCI oligomer DNA for use in a ligation reaction. One pmole of ds NotIBfuCI oligomer was ligated to 10 mL of the re-dissolved Ac-DAF-16 DBD bound DNA with T4 DNA ligase in 20 mL total volume. Two microliters of the ligation product were used as template in PCR reactions with NotIBfuCI PCR primer (59-AAAAGGGGCGGCCGCG ATC-39). The cycling conditions were 4 min at 95uC, followed by 29 cycles of 1 min at 95uC, 1 min at 57uC, 2 min at 72uC and a final extension for 6 min at 72uC.
Amplified PCR products were purified with NucleoSpin Extract II kit (Clontech). Two micrograms of the DNA were used for subsequent rounds of binding experiments, as described above. After each round of binding and elution, PCR was performed for a total of 4 rounds. Purified PCR products from the last round were subcloned either into the NotI site of pBluescript KS+ vector (Stratagene) or directly into pGEM -T Easy vector (Promega) for sequencing and analysis.

Motif analysis of hookworm DAF-16 bound genomic fragment
Constructs containing Ac-DAF-16 bound fragments from the in vitro genomic selection cloned in pBluescript KS+/pGEM-T were sequenced with vector specific primer T7. The sequence chromatograms were screened using sequence display software Chromas 2.33 (Technelysium Ltd.) to identify high quality regions, vector linker regions, and the low complexity regions. All subsequent analyses were based on the high-quality sequences with vector linker sequences removed.
Nucleotide sequences were searched against the draft A. caninum genome scaffold sequences [39] (Mitreva, unpublished) using BLAST to confirm the identity of the hookworm genomic fragments. If the overall percent identity of an alignment was above 95%, the genome scaffold was assumed to be a hit. The bound fragments that failed to hit any A. caninum genome scaffolds were further searched against nucleotide databases at GenBank (http://www.ncbi.nlm.nih. gov/) to detect possible contamination from other genomes. The scaffold hits were next analyzed for their coding potential by mapping the cDNAs to the genome sequence (see below).
Given the current knowledge of reported FOXO transcription factor binding sites, Ac-DAF-16 bound A. caninum genome fragments were first inspected for existence of the consensus DBE, 59-TTG/ATTTAC-39 (and the corresponding reverse compliment 59-GTAAAC/TAA -39) [19], or its reverse counterpart, 59-CATTTA/GTT-39 (and the corresponding reverse compliment 59-AAT/CAAATG-39). The motif discovery algorithm Gibbs Motif Sampler [50,51] was applied to sequences lacking canonical binding sites to seek over-represented motifs as additional candidate Ac-DAF-16 response elements. The default eukaryotic parameters were used to run the Gibbs Motif Sampler program with a slight modification made to the target motif widths, with 6, 8, 10, 12, and 14 bp chosen based on the width range of C. elegans transcription factor binding sites in Open REGulatory ANNOtation database (ORegAnno) [52,53]. Two other motif discovery algorithms, MEME [54] and Mdscan [55], were also attempted for de novo motif discovery; however, unlike the robust Gibbs Motif Sampler, either their performance in detecting the consensus DBE with a set of positive control sequences was poor or the parameter setting was not flexible. Therefore, only Gibbs Motif Sampler was used for the current data set. Returned motifs were ranked based on the occurrence rates in the bound genomic fragments. Sequence logos for identified motifs were generated using WebLogo [56]. All of the cDNA and genomic sequences are available at the Nematode.net FTP site (http://nematode.net/FTP/index.php) [57].

Identification of potential Ac-DAF-16 regulated genes
Ancylostoma caninum genome scaffold sequences containing Ac-DAF-16 bound genomic fragments, stage specific A. caninum expressed sequence tag (EST) databases [58], and the recent A. caninum cADNA databases [31] were obtained from Nematode.net [57]. The available data sets cover 93% of the A. caninum transcriptome.
For each genome scaffold hit, the genomic template including the Ac-DAF-16 bound region and up to 3 kb extensions at both ends were extracted. Those templates were searched against EST databases/transcriptome databases using BLAST to identify corresponding transcripts. The cut-off for the overall percent identity of an alignment was 93% and a length of at least a 100 bp. Gene location was predicted with greater confidence if more ESTs were aligned. Gaps between two adjacent exon segments in an EST alignment were treated as possible introns and were confirmed by checking intron boundary sequences (GT/AG rule) [59]. The transcript contigs were then derived from those ESTs.
The identified transcripts were compared with existing protein sequences at GenBank for functional annotation, using a maximum E-value of 1610 210 and a minimum of 50% similarity as cut-off. Their expression specificity across different developmental stages (L3, aL3, F and M) were characterized using a statistical approach defined by Audic [31,60] Briefly, the cDNAs were organized into transcript contigs, and ESTs were grouped to the corresponding cDNA libraries derived from different developmental stages for each transcript contig. The frequencies of library specific cDNAs for each contig were recorded and analyzed using a modified Fisher's exact test with a significance of p,1 e-05. This allowed definition of the stage specificity of transcript expression with significance. Transcripts were also mapped to the three organizing principles of the Gene Ontology (GO) based on sequence similarity displayed using AmiGo (http://amigo. geneontology.org), and are available at Nematode.net (http:// nematode.net).